The public health implications of reliable risk prediction tools for estimating the probability of lung cancer, the leading cause of cancer mortality in the US, are immense. In this proposal, we build upon an epidemiologic risk model for lung cancer that we recently developed from a lung cancer case-control study of 1851 Caucasian lung cancer patients and 2001 controls, matched to the cases on sex, age ([unreadable]5 years), smoking status (current, former, never) and ethnicity from the parent grant (ROl CA55679, P1: M. Spitz). This risk model also estimates an individual's absolute risk for lung cancer. Using an independent data set from this study and an external data set of lung cancer cases and controls from Dr. David Christiani (co-investigator, Harvard School of Public Health, CA74386), we propose to validate the lung cancer risk model and extend it by incorporating a genetic biomarker of risk. Specifically: 1). In 800 prospectively accrued cases and 800controls using the recruitment mechanisms of parent grant, we will validate our original model and then assess the added discriminatory ability of a promising novel cytogenetic biomarker, the cytokinesis-block micronucleus (CBMN) assay, a multi-endpoint assay that measures not only chromosome damage (micronuclei reflecting chromosome breaks;nucleoplasmic bridges reflecting chromosome rearrangements and nuclear buds reflecting gene amplification) but also other cellular events (apoptosis and necrosis). We will measure these endpoints at baseline and following challenge with the tobacco-specific nitrosamine, NNK. We will derive a method to integrate these different measures of chromosome/genetic instability into the epidemiologic lung cancer risk model. 2). Using the findings from aim 1, we will construct an extended risk model to include measures of chromosome instability and gene-environment interactions. Our preliminary data show that our current model has moderate discriminatory power (70%), we believe that extending the model to includes these biomarkers of chromosome instability as well as gene environment interactions will only improve the discriminatory power of our model. This newly developed model may be useful to identify high-risk populations who could then be targeted for intensive smoking-cessation programs and could be enrolled into chemopreven-tion screening trials. 3). Internally validate the original and extended lung cancer risk models using an additional set of 500 prospectively enrolled lung cancer cases and 500 controls using the recruitment mechanisms of parent grant and compare the discriminatory power between the extended model to that of the original Spitz model between the two models. This will also include independent, internal, validation of the CBMN assay. Evaluation of these chromosomal endpoints in an independent sample will provide proof-of-principle for subsequent inclusion of additional functional phenotypes and genotypes into the model.